Synthase from Euglena gracilis'

Transcription

1 Plant Physiol. (1991) 97, /91/97/1271 /9/$1./ Received for publication May 14, 1991 Accepted July 16, 1991 Purification and Characterization of Chorismate Synthase from Euglena gracilis' Comparison with Chorismate Synthases of Plant and Microbial Origin Andreas Schaller, Manfred van Afferden2, Volker Windhofer3, Sven Bulow4, Gernot Abel, Jurg Schmid, and Nikolaus Amrhein* Institute of Plant Sciences, Swiss Federal Institute of Technology, Zurich, Switzerland (A.S., G.A., J.S., N.A.); and Lehrstuhl fur Pflanzenphysiologie, Ruhr-Universitat Bochum, Federal Republic of Germany (M. v.a., V. W., S.B.) ABSTRACT Chorismate synthase was purified 12-fold from Euglena gracilis. The molecular mass of the native enzyme is in the range of 11 to 138 kilodaltons as judged by gel filtration. The molecular mass of the subunit was determined to be 41.7 kilodaltons by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Purified chorismate synthase is associated with an NADPH-dependent flavin mononucleotide reductase that provides in vivo the reduced flavin necessary for catalytic activity. In vitro, flavin reduction can be mediated by either dithionite or light. The enzyme obtained from E. gracilis was compared with chorismate synthases purified from a higher plant (Corydalis sempervirens), a bacterium (Escherichia co/i), and a fungus (Neurospora crassa). These four chorismate synthases were found to be very similar in terms of cofactor specificity, kinetic properties, isoelectric points, and ph optima. All four enzymes react with polyclonal antisera directed against chorismate synthases from C. sempervirens and E. coli. The closely associated flavin mononucleotide reductase that is present in chorismate synthase preparations from E. gracilis and N. crassa is the main difference between those synthases and the monofunctional enzymes from C. sempervirens and E. coli. The three aromatic amino acids are synthesized via the shikimate pathway in microorganisms and plants (Fig. 1). Although the reaction sequence in the prechorismate pathway is identical in all organisms investigated so far, there are considerable differences in enzyme organization, as well as regulation, between organisms of different taxonomic groups. In Escherichia coli, the seven enzymatic activities of the prechorismate pathway are associated with monofunctional polypeptides. In contrast, in fungi the five enzyme activities ' Supported at Ruhr-Universitat Bochum by a grant from the Deutsche Forschungsgemeinschaft and at the Swiss Federal Institute of Technology by ETHZ grant 7592/ Present address: DMT Gesellschaft fuir Forschung und Prufung mbh, Essen, FRG. 'Present address: Institut fur physiologische Chemie, Ruhr Universitit Bochum, FRG. 4Present address: PALL Biomedizin GmbH, Dreieich, FRG converting DAHP5 to EPSP reside in a single pentafunctional polypeptide, the arom complex. In higher plants, two activities (3-dehydroquinate dehydratase and shikimate dehydrogenase) form a bifunctional polypeptide, although the other five enzymes are monofunctional (reviewed in ref. 4). In Euglena gracilis, the organization of the prechorismate pathway appears to resemble that found in fungi. The first and the last steps are catalyzed by single enzymes (DAHP-synthase and chorismate synthase, respectively), whereas activities 2 to 6 form a large complex resembling the fungal arom complex (27). Euglena and fungi are strikingly similar in other biochemical pathways as well. Thus, both groups of organisms use the L-a-aminoadipate pathway for lysine biosynthesis, whereas bacteria, algae, and plants follow the diaminopimelate route (3). On the basis of such findings, a close evolutionary relationship was proposed for euglenoids and fungi. But aromatic biosynthesis shows several features that, among all organisms studied so far, are unique for E. gracilis. Anthranilate synthase, the first enzyme in the tryptophan branch (Fig. 1), is monomeric in Euglena, whereas in all other anthranilate synthases characterized so far the chorismate and glutamine binding sites reside on distinct polypeptide chains (16). The synthesis of tryptophan from anthranilate is catalyzed by a single protein in Euglena (17), but in all other known examples tryptophan synthase is separable from the other activities of the tryptophan branch (17). Unlike other organisms, E. gracilis possesses only a single chorismate mutase isozyme, which is unusually large (31). In contrast to microorganisms, L-arogenate is the last common precursor of both phenylalanine and tyrosine in Euglena (7), which also appears to be the case for higher plants. Chorismate synthase is the only shikimate pathway enzyme previously not detected in E. gracilis, except indirectly (6), which remained to be characterized. A comparison of the Euglena enzyme with chorismate synthases of other species is particularly interesting considering the data mentioned above. Chorismate synthase (EC ) catalyzes the formation 5Abbreviations: DAHP, 3-deoxy-D-arabino-heptulosonate 7-phosphate; EPSP, 5-enolpyruvylshikimate 3-phosphate; FMN, flavin mononucleotide; FAD, flavin adenine dinucleotide.

2 1 272 SCHALLER ET AL. Plant Physiol. Vol. 97, 1991 E4P Q + ' DAHP 'DHQ PEP SHK ' S3P '- EPSP ' ' (is DHS- CHA \@1 1. DAHP-Synthase 2. DHQ-Synthase PPA 3. DHQ-Dehydratase j1\ 4. SHK-Dehydrogenase 5. SHK-Kinase I 6. EPSP-Synthase AGN 7. CHA-Synthase 8. CHA-Mutase I, 9. ANT-Synthase Pthe tyr trp ANT Figure 1. Overview of the shikimate pathway. E4P, D-erythrose 4- phosphate; PEP, phosphoenolpyruvate; DHQ, 3-dehydroquinate; DHS, 3-dehydroshikimate; SHK, shikimate; S3P, shikimate 3-phosphate; CHA, chorismate, PPA, prephenate; AGN, arogenate; ANT, anthranilate. of chorismate-the last common precursor for all three aromatic amino acids-by 1,4-elimination of orthophosphate from EPSP. The enzyme strictly requires a reduced flavin as cofactor, although no net change in redox state ofthe substrate is observed. Chorismate synthases have been purified and characterized from E. coli (25, 33), Neurospora crassa (32, 33), and Bacillus subtilis (18). Although chorismate synthase activity was only recently discovered in an extract from a higher plant (26), the enzyme has been isolated and characterized from a cell suspension culture of Corydalis sempervirens in our laboratory (28). Furthermore, a cdna coding for C. sempervirens chorismate synthase has been cloned and analyzed (A. Schaller, J. Schmid, U. Leibinger, N. Amrhein, submitted for publication). Here we report the purification and characterization of chorismate synthase from E. gracilis as a representative of lower photoautotrophic eukaryotes, together with a more detailed analysis of the properties of the C. sempervirens chorismate synthase. The results are discussed in the context of data obtained for chorismate synthases isolated from E. coli and N. crassa. We find that chorismate synthase, unlike other shikimate pathway enzymes, is highly similar in different organisms. The main difference appears to be the presence of a flavin reductase activity that is associated with chorismate synthase in some fungi and E. gracilis. Strains and Cultures MATERIALS AND METHODS Euglena gracilis strain /25 was obtained from the German Collection of Microorganisms (Gottingen, FRG). It was grown in NL medium (5) either on plates or in liquid cultures of 25, 1, or 75 ml in Erlenmeyer flasks on a rotary shaker at 25 C under constant illumination with incandescent white light (1 ME m-2s-1). The 75 ml cultures were used to inoculate a 2 L culture in an airlift fermenter. Cells were harvested after 7 d (late logarithmic phase of growth) at a cell density of 2 x 16/mL. A cell suspension culture of C. sempervirens was grown as described previously (28). The E. coli strain AB2849 carrying the plasmid pgm62 (33) was a gift of Dr. J.R. Coggins (University of Glasgow, UK). It was grown in 2YT medium in the presence of ampicillin. Materials All chemicals purchased from commercial sources were of the highest purity available. Chorismate was isolated and purified according to the method of Gibson (15). Shikimate 3-phosphate was prepared as in ref. 21 and EPSP was synthesized from shikimate 3-phosphate as described in ref. 22. Anthranilate synthase from Aerobacter aerogenes was purified following the procedure of Egan and Gibson (11) with the modifications described in ref. 28. Chorismate synthase was purified from C. sempervirens as described previously (28). Chorismate synthases were purified from E. coli about 4- fold and from N. crassa 26-fold essentially as described in ref. 33 with omission ofthe last chromatographic step. Specific enzyme activities were 48 and 6 nkat/mg, respectively. An antiserum was raised in rabbits against C. sempervirens chorismate synthase (28) and the immunoglobulin G fraction was purified on protein A beads (Pharmacia). The rabbit anti-e. coli chorismate synthase serum was a gift of Dr. J.R. Coggins. Assay of Chorismate Synthase Activity All enzyme reactions were at 3 C unless otherwise stated. Chorismate synthase activity can be assayed either by detection of the reaction products (chorismate or phosphate) or by forward coupling to the reaction of anthranilate synthase and monitoring the anthranilate production. Throughout the enzyme purification, the discontinuous coupled assay was used, according to ref. 32 for E. gracilis and N. crassa, and according to ref. 28 for C. sempervirens and E. coli. The coupled assay was not used for the characterization of chorismate synthases, so as to exclude the influence of anthranilate synthase on the results. For characterization of the E. gracilis chorismate synthase, the formation of chorismate was followed at 274 nm using the assay system described in ref. 33. The assay mixture contained 5 mm triethanolamine/hcl, ph 8., 5 mm KCI, 2.5 mm MgCl2, 83 um NADPH, 1 Mm F-MN, and 8 um EPSP. This assay is not compatible with chorismate synthases that lack an intrinsic flavin reductase activity. When measuring the E. coli and C. sempervirens chorismate synthase activities, the flavin had to be reduced either chemically (5 mm dithionite) or by illumination (white incandescent light, 75 MuE m-2s-'). Both treatments interfered, however, with the spectrophotometric assay of chorismate formation. Therefore, the rate of phosphate release was determined for the characterization of the plant enzyme as well as for the comparative studies with all four chorismate synthases. The reaction mixture (total volume 25 ul) contained 5 mm triethanolamine/hcl, ph 8., 1,uM FMN, 2 mm DTT, 25 mm KCI, 32,uM EPSP. The reaction was initiated after 1 min illumination (to reduce the FMN) by addition of EPSP. After 15 or 3 min under continuous white

3 CHARACTERIZATION OF CHORISMATE SYNTHASE light (75,uE m-2s-'), free Pi was determined according to Lanzetta et al. (24). Purification of Chorismate Synthase from Euglena gracilis All steps were carried out at 4 C unless otherwise stated. 1. Protein Extraction and Precipitation E. gracilis cells were harvested by centrifugation and resuspended in an equal volume of 1 mm Tris/HCl, ph 7.5, 5 mm EDTA, 1.2 mm PMSF, 1 mm DTT, and disintegrated by sonication. Cell debris was removed by centrifugation (3 min, 27, g). The resulting supernatant, designated crude extract, was subjected to fractionated ammonium sulfate precipitation from 35 to 65% saturation. The precipitate was collected by centrifugation (as above) and resuspended in a small volume of 4 mm Tris/HCl, ph 7.5, 5 mm EDTA, 1.2 mm PMSF, 1 mm DTT (buffer A). Then the solution was desalted by chromatography on Sephadex G-25 (Pharmacia, equilibrated in buffer A). 2. DEAE-Sepharose Chromatography The desalted protein solution was loaded onto a DEAE- Sepharose (Pharmacia) column (1 ml bed volume) preequilibrated in buffer A. The column was first washed with buffer A, then protein was eluted with a linear gradient (5 ml) of 1 to 27 mm KCl in buffer A (flow rate 1 ml/min, 1 ml fractions). Fractions containing chorismate synthase activity were pooled. 3. Phenyl-Sepharose Chromatography The combined fractions from the previous step were loaded directly onto a Phenyl-Sepharose column (Pharmacia, 4 ml bed volume), equilibrated in buffer A. Protein was eluted with a gradient (4 ml) of to 3% (v/v) ethyleneglycol (flow rate.8 ml/min, 1 ml fractions). Fractions containing chorismate synthase activity were pooled and concentrated by ultrafiltration (Amicon PM-3). Glycerol and DTT were added to a final concentration of 5% (v/v) and 1 mm, respectively. This preparation could then be stored for several months without any loss of chorismate synthase activity. 4. MonoQ Chromatography Chromatography on a MonoQ HR 5/5 Fast Protein Liquid Chromatography column (Pharmacia) was used as final step in chorismate synthase purification when the purified enzyme was to be used for biochemical characterization because of the high yield and speed of the method. The protein sample containing 5% (v/v) glycerol was applied to the column that had been preequilibrated in 2 mm Tris/HCl, ph 7.5 (buffer B). Protein was eluted using first a gradient (14 ml) of to 23 mm KCl in buffer B, then a second gradient (8 ml) of 23 to 275 mm KCl in buffer B (flow rate 1 ml/min, 1 ml fractions). Combined fractions with chorismate synthase activity were desalted on Sephadex G-25 equilibrated in buffer B containing 1 mm DTT. 5. Cellulose Phosphate Chromatography In some cases, chromatography on cellulose phosphate (P1 1, Whatman) was used as the last purification step, replacing MonoQ chromatography. This step was more time consuming and gave a lower yield but resulted in a much higher purification and, therefore, was used when chorismate synthase was intended to be analyzed by PAGE. Ammonium sulfate was added to the combined fractions of step 3 up to 8% saturation. Precipitated protein was collected by centrifugation and resuspended in a small volume of 1 mm potassium phosphate buffer, ph 7.5, 5 mm EDTA, 1 mm DTT (buffer C), and desalted on Sephadex G-25 equilibrated in the same buffer. The sample was then loaded onto a cellulose phosphate column (2 ml bed volume) that had been preequilibrated in buffer C. The column was washed with 1 mm phosphate in buffer C; then chorismate synthase was eluted with 2 mm phosphate in buffer C. The eluate was dialyzed against 4 mm Tris/HCl, ph 7.5, 5 mm EDTA, 1 mm DTT, 5% glycerol, and subsequently stored at -2 C. Determination of Molecular Mass and of Isoelectric Point The molecular mass of native Euglena chorismate synthase was determined by gel filtration on an Ultropac TSK G3SW HPLC column (LKB) preequilibrated with 2 mm Tris/HCl, ph 7.5, 1 mm KCl, at a flow rate of.5 ml/ min. The molecular mass of chorismate synthase was determined in comparison with those of standard proteins (alcohol dehydrogenase, 15 kd; ovalbumin, 45 kd; carboanhydrase, 29 kd; Cyt c, 12.4 kd). The isoelectric point was determined by chromatofocusing on PBE 94 (Pharmacia, 11 ml bed volume), equilibrated in 25 mm imidazol/hcl, ph 7.4 or ph 6.8, 1 mm DTT, for chorismate synthases of E. gracilis and N. crassa, respectively. The protein sample was dialyzed against equilibration buffer prior to application. Protein was eluted by a linear ph gradient created by Polybuffer 74 (Pharmacia)/HCl, ph 4., 1 mm DTT, with a flow rate of 1 ml/min. Polybuffer 74 was used in eightfold dilution. PAGE and Westem Blot SDS-PAGE was performed on 12.5% polyacrylamide slab gels using the buffer system described by Laemmli (23). Gels were silver stained after electrophoresis according to Ansorge (1). Alternatively, proteins were transferred to nitrocellulose membranes using a semidry blotting apparatus (Millipore). Immunodetection of proteins on the blots was done at room temperature with 3% BSA in Tris-buffered saline as blocking agent. Goat anti-rabbit immunoglobulin G alkaline phosphatase conjugate obtained from Bio-Rad was used as secondary antibody. Protein Concentration Protein concentration was determined using the Bio-Rad protein assay following the manufacturer's instructions with BSA as a standard.

4 1 274 SCHALLER ET AL. Plant Physiol. Vol. 97, 1991 RESULTS AND DISCUSSION Purification of Chorismate Synthase from E. gracilis The purification of chorismate synthase starting from 5 g (wet weight) of E. gracilis cells is summarized in Table I. Chorismate synthase activity was not readily detectable in crude extracts of E. gracilis due to the presence of low molecular weight inhibitory material. To determine chorismate synthase activity in the crude extract, this material could be removed by gel filtration and it was lost during ammonium sulfate precipitation. Chorismate synthase activity eluted at a conductivity of 9.7 msiemens/cm from DEAE-Sepharose, well separated from the bulk protein. Fractions containing chorismate synthase activity could then be applied directly to Phenyl-Sepharose. The major portion of proteins did not bind to the column. Chorismate synthase activity eluted at 5% (v/ v) ethyleneglycol. Chromatography on MonoQ was the final step during routine enzyme preparation, resulting in a 269- fold enrichment of chorismate synthase with a yield of 54%. In this preparation, enzyme activity was linear both with time and with protein concentration in the enzyme assay, and therefore it was considered suitable for enzyme characterization. A better, about 12-fold purification could be achieved at expense of the yield when chromatography on cellulose phosphate was performed as the final step in the purification procedure. But even this highly purified chorismate synthase preparation was not homogeneous as judged by SDS-PAGE (Fig. 2). There are still at least four different protein bands visible. One of them, with a Mr of 41,7, reacts with antisera directed against chorismate syntheses of both E. coli and C. sempervirens (Fig. 3) and could thus be identified as E. gracilis chorismate synthase. The apparent molecular mass of the native enzyme was determined to be in the range of 11 to 138 kd by gel filtration (data not shown). The 41.7 kd polypeptide, therefore, is a subunit of native chorismate synthase. Chorismate syntheses of other organisms are also composed of multiple subunits. White et al. (33) reported that chorismate syntheses of E. coli and N. crassa are both homotetramers. Gaertner (13) observed at least two active forms of the enzyme in N. crassa, the homodimer and the homotetramer. In C. sempervirens, native chorismate synthase consists of at least two identical subunits (28). Native chorismate synthase in Bacillus subtilis is composed ofthree polypeptides carrying chorismate synthase, 3-dehydroquinate synthase, and a flavin reductase activity, respectively (18). The apparent molecular mass of native E. gracilis chorismate synthase Table 1. Chorismate Synthase Purification from E. gracilis Cells (5 g) Purification Stage Total Specific Recovery Purfication Activity Activity ReoryPifctn nkat nkat/mg % -fold Crude extract (NH4)2SO4 precipitate DEAE-Sepharose Phenyl-Sepharose MonoO Cellulose-phosphate '`.. 3.E ". --- BP Figure 2. Silver-stained SDS-PAGE documenting the progress of chorismate synthase purification from E. gracilis. Lanes 1 and 5, protein standards (Low Molecular Weight Calibration Kit, Pharmacia). Lanes 2 to 4, combined fractions containing chorismate synthase activity after chromatography on DEAE-Sepharose (12 Og of protein), Phenyl-Sepharose (6 Mg), and cellulose phosphate (.8 Ag), respectively. suggests either a homotrimeric structure of three 41.7 kd polypeptides or an oligomeric structure of different subunits. The latter possibility is supported by the observation that the Mr of 41,7 is similar to that of monofunctional chorismate syntheses (41,9 and 39,138 in C. sempervirens [28] and E. coli [8], respectively). The flavin reductase activity, which is still present in the highly purified Euglena enzyme, may therefore be located on a different subunit. Cofactor Requirement of Chorismate Synthases Chorismate synthase strictly requires a reduced flavin cofactor for enzymatic activity. For in vitro analysis, the reduced flavin has either to be supplied exogenously, as in E. coli (33) and C. sempervirens (28), or it can be generated by a flavin reductase activity that may be associated either covalently or noncovalently with chorismate synthase as described for N. crassa (32, 33) and B. subtilis (18), respectively. In E. gracilis,

5 CHARACTERIZATION OF CHORISMATE SYNTHASE 1275 C. sempervirens E. gracills E. coi N. crassa M k - MOON 67 k -.me 43 k k,,: f.. z: X 4::j. :,.jjj...z.:.;,. -:...".-. 1:.-lik"I"..":.1, 4i ,-,,.nmwmv 3 k- Figure 3. Immunological cross-reactivity of chorismate syntheses from different organisms. Lanes 1: silver-stained SDS-PAGEs of chorismate synthase purified from C. sempervirens (2 pkat/.5,g), E. graclils (15 pkat/1 Atg), E. coli (25 pkat/.5 Mg), and N. crassa (12 pkat/5 Mg). Lanes 2 and 3: proteins detected on Western blots by polyclonal antisera directed against chorismate synthase from C. sempervirens and E. coli, respectively. Compared with lane 1, four times as much protein was loaded in lanes 2 and 3. Only in C. sempervirens (lane 2) and E. coli (lane 3), where homologous antisera were used, the protein amount was the same as in lanes 1. Lane M: protein standards (Low Molecular Weight Calibration Kit, Pharmacia). there is a flavin reductase activity associated even with the highly purified chorismate synthase, which strictly depends on NADPH as a source of electrons for FMN reduction (data not shown). The four chorismate syntheses that are the subject of the present investigation all show a very similar behavior when supplied with flavin derivatives exogenously reduced by 5 mm dithionite. Chorismate synthase activity is stimulated most effectively by FMN, which can be partially replaced by FAD and, to a lesser extent, also by riboflavin (Table II). When E. gracilis chorismate synthase is assayed with NADPH instead of dithionite as a reducing agent, i.e. by using the intrinsic flavin reductase activity, the specificity for FMN is much more pronounced. Under these conditions, FAD and riboflavin do not stimulate chorismate synthase to more than background activity (without addition of any flavin). We conclude from these data that the FMN specificity of chorismate synthase from E. gracilis reflects in fact the specificity of the flavin reductase. This was also reported for the bifunctional N. crassa enzyme by White et al. (33). In contrast with their observation, we could also detect activity of N. crassa chorismate synthase in the presence of only NADPH and FAD. This result may well be explained by the fact that our enzyme preparation was less pure than that used by White et al. (33), and that, thus, a second contaminating flavin reductase may have been present in our preparation that reduced FAD at the expense of NADPH. The flavin reductase that is noncovalently associated with chorismate synthase in B. subtilis accepts FMN and FAD as equivalent electron donors (18). B. subtilis chorismate synthase has so far not been assayed independently of its flavin reductase. Light Stimulation of Chorismate Synthase Activity The reduction of a site in chorismate synthase, which has been suggested to be either Fe(III) (25) or a disulfide bond (32), appears to be a prerequisite for maintaining the active state of the enzyme. FMN appears to be involved in this process in conjunction with a flavin reductase activity. The flavin reductase is associated with chorismate synthase in N. crassa, B. subtilis, and E. gracilis. In E. coli, FMN is presumably reduced by an endogenous diaphorase activity (33). The in vivo source of reduced flavin for plant chorismate syntheses is unknown. Because plant chorismate synthase activity is associated with the chloroplast compartment, a possible involvement of a photosynthetic electron transport chain component has been discussed (26). This would allow for the regulation of the shikimate pathway by light. A light stimulation of the overall activity of the shikimate pathway, in vivo or in isolated chloroplasts, has been reported earlier ( 19, 2), but light-regulated enzymes have not been identified yet (29). We found that light (white incandescent light, 75 ME m-2s-'), in the presence of 1 mm DTT, can provoke chorismate synthase activity in the absence of a flavin reductant (Table III). This light stimulation of enzyme activity is most likely not a direct effect of light on the enzyme itself, but is rather due to photoreduction of FMN. We observed that only light

6 1 276 SCHALLER ET AL. Plant Physiol. Vol. 97, 1991 Table II. Chorismate Synthase Activity as a Function of Flavin Cofactors and Reducing Agents Chorismate synthases purified from E. gracilis, N. crassa, C. sempervirens, and E. co/i were assayed in the presence of different flavin cofactors (1 AM each). The concentration of dithionite was 5 mm and that of NAD(P)H 2 Mm. Activities in each column are expressed as percentages of the activities in the presence of FMN. All experiments were done in triplicate. Chorismate Synthase Activity Flavin E. gracilis N. crassa C. sempervirens E. coli NADPH Dithionite NADPH Dithionite NADPH Dithionite NADH Dithionite FMN 1 oa 1 oob 1ooC 1 ood e loot e 19 FAD Riboflavin None a85 1% is equivalent to a 13, b 27, c 255, d 381, e not detectable, '364, and pkat/ml. qualities capable of FMN photoreduction (i.e. blue light) could also promote chorismate synthase activity. Permanent illumination, which keeps the FMN in its reduced state and thus produces essentially anaerobic conditions, is necessary for maintenance of chorismate synthase activity (data not shown). Whether or not this effect is of any physiological relevance is not clear. The fact that E. coli chorismate synthase also is stimulated by light to some extent may argue against a direct physiological role of light in the activation of chorismate synthase. On the other hand, free flavins have been proposed to act as primary photoreceptors in the photoreactivation of N. crassa nitrate reductase (12). On the basis of our data, we cannot clearly assign to chorismate synthase a function in the regulation of the shikimate pathway. Using light as reducing power, it was now possible to assay the activity of the monofunctional chorismate synthases without forward coupling of the reaction to anthranilate synthase (see "Materials and Methods"). This assay system, which dispenses with the highly tryptophan-sensitive anthranilate synthase, allowed us to test a possible influence of aromatic amino acids on chorismate synthase activity. In isolated spinach chloroplasts, each of the aromatic amino acids was shown to exert strict feedback inhibition of its own synthesis (2). Chorismate synthase activity, however, was not affected by end products ofthe pathway (Table IV). None of the aromatic Table Ill. Stimulation of Chorismate Synthase Activity by Light Chorismate synthases purified from E. gracilis, N. crassa, C. sempervirens, and E. coli were assayed under continuous illumination (75 ME m-2 s-' white incandescent light), which could partially replace the reducing agents dithionite (5 mm) or NADPH (2 PM). The activity is given as percentages of the value obtained in the presence of 5 mm dithionite. All experiments were done in triplicate. Reducing Agent Chorismate Synthase Activity E. gracilis N. crassa C. semperivirens E. coli Dithionite, dark 18 1 oob 1 QOC 1 ood NADPH, dark None, dark None, light ad 1% is equivalent to 272, b 387, c 328, d 445 pkat/ml. amino acids tryptophan, phenylalanine, tyrosine, or anthranilate (1 mm each) had a significant effect on chorismate synthase activity. On the other hand, the reaction product chorismate severely reduced the activity of all four chorismate synthases at a concentration of mm. Kinetic Properties Apparent Km values were deduced from steady-state kinetic analyses. In all cases, triethanolamine/hcl buffer was used at a ph of 8., which is close to the ph optima of the known chorismate synthases (8., 8.2, and 8.1 for C. sempervirens, E. gracilis, and N. crassa, respectively). Figure 4 shows the Michaelis-Menten and linearized plots from which the apparent Km values were calculated for the C. sempervirens and E. gracilis enzymes. The apparent Km values for EPSP were 53 and 27 MM, and those for FMN were 37 and 76 nm for the C. sempervirens and E. gracilis chorismate synthases, respectively. The N. crassa enzyme has a comparably high affinity for FMN (data not shown), whereas the affinity of the B. subtilis chorismate synthase (18) is much lower (66 nm and 12.5 MM, respectively). An apparent Km for EPSP of 2.7 MM was reported earlier (3) for N. crassa chorismate synthase, although we calculated a value of 7 gm. With the exception Table IV. Effect of Aromatic Amino Acids on Chorismate Synthase Activity Chorismate synthases purified from E. gracilis, N. crassa, C. sempervirens, and E. coli were assayed in the presence of aromatic amino acids and chorismate (1 mm each). The activity is given as percentages of the value obtained in the photochemical assay without addition of any effector. All experiments were done in triplicate. Chonsmate Synthase Activity Amino Acid E. gracilis N. crassa C. sempervirens E. coli Tryptophan 1 5a 1 8b 1 8c 1 8d Phenylalanine Tyrosine Anthranilate Chorismate a-d 1% is equivalent to a871, b 826, c 167, d 972 pkat/ml.

7 CHARACTERIZATION OF CHORISMATE SYNTHASE B 15- E 6 'EL E 1 [EPSPJ ( p M) [EPSP] (p M) 5 C 2. D 4 a 15 F A 16 E 4 Ẹa 1 Q - 2 E B Ec 8 >2 z. 5 [ I ljmni 1, M) o , I FtN ", LFMNI p M) (FMNJ ( M) Figure 4. Determination of apparent Km values for EPSP and FMN. Enzyme activity is plotted versus EPSP and FMN concentrations for C. sempervirens (A, C) and E. gracilis (B, D) chorismate synthases, respectively. The inserts show the corresponding linearized plots, which were used for determination of the apparent Km values. of the B. subtilis enzyme, chorismate synthases show similar affinities for EPSP and FMN in all organisms from which data are available. High concentrations of FMN were inhibitory for all chorismate synthases investigated here. An inhibitory effect of high EPSP concentrations was observed for E. gracilis and N. crassa chorismate synthases. Immunological Relationships between Chorismate Synthases Polyclonal antisera directed against the chorismate synthases of C. sempervirens and E. coli were used to detect chorismate synthases of different species on Western blots (Fig. 3). Both antisera react with chorismate synthase of E. gracilis, as mentioned above, but they also recognize the other chorismate synthases under investigation. Chorismate synthases, or rather the domains of the polypeptides carrying chorismate synthase activity (in the case of bifunctional enzymes), therefore seem to be highly similar in organisms that are widely apart in evolutionary terms. Whereas the interpretation in quantitative terms of differing intensities of stained protein bands may be deceiving, we wish to point out that the antiserum directed against plant chorismate synthase gives bands of equal intensity with the three other chorismate synthases, although the amount of chorismate synthase protein present on the gels seems to be by far the smallest in case of E. gracilis as judged by the silver-stained lanes. This may indicate a closer relationship between the chorismate synthases of Corydalis and Euglena on the one hand than between the Corydalis enzyme and the bacterial and fungal chorismate synthases on the other hand. It is at present not possible to specify the immunological relationship more precisely because we cannot be sure of the amount of antigen present in each reaction. Similar amounts were loaded onto

8 1 278 SCHALLER ET AL. Plant Physiol. Vol. 97, 1991 Table V. Synopsis of Chorismate Synthase Properties The data presented in the "Results and Discussion" section are summarized. Numbers in parentheses indicate the reference for data published elsewhere. E. gracilis N. crassa C. sempervirens E. coli Apparent Km (EPSP) 27,M 7 (2.7) ZlM (3) 53 um n.d.a Apparent Km (FMN) 76 nm 66 nm 37 nm n.d. Cofactor specificity FMN > FAD FMN > FAD FMN > FAD FMN > FAD Flavin reductase Present Present(32) Absent (28) Absent (33) Apparent Km (NADPH) 7.4 itm 21.4 AM Mol. mass (native) kd 198 kd (33) 8.1 kd (28) 144 kd (33) Mol. mass (denatured) 41.7 kd 5 kd (33) 41.9 kd (28) 39.1 kd (8) No. of subunits n.d. 2 or 4 (13, 33) 2 (28) 4 (33) Isoelectric point (28) n.d. ph optimum (25) Not determined. the gel in terms of enzyme activity, but the specific activities of the preparations differed. Also, the intensity of the silver stain may not accurately reflect the actual amount of protein present, because not all proteins stain equally in this procedure. The immunological cross-reactivity of all four chorismate synthases indicates that antigenic determinants ofthis enzyme were highly conserved during evolution. Also, the primary structures of chorismate synthases turn out to be highly similar throughout species. There is 48% sequence identity between the enzymes from C. sempervirens and E. coli (A. Schaller, J. Schmid, U. Leibinger, N. Amrhein, submitted for publication) and 53% between C. sempervirens and Saccharomyces cerevisiae (D. Jones, G. Braus, personal communication). EPSP synthase appears to be the only other enzyme in the prechorismate pathway that shows a similar degree of conservation (54% identity between EPSP-synthases of petunia and E. coli [14]). In contrast, the individual domains of the arom complex from S. cerevisiae are only between 21 and 38% identical to the respective corresponding monofunctional E. coli enzymes (9), and potato DAHPsynthase shows only as little as 22% identity with the three E. coli DAHP-synthases (1). Clearly, until sequences of other enzymes of the pathway become available for comparison, it is too early for a generalization. CONCLUSIONS We have shown here that chorismate synthases from E. gracilis, C. sempervirens, N. crassa, and E. coli, i.e. of species representing four distinct taxonomic phyla, must be considered very similar proteins (see synopsis in Table V). The enzyme from B. subtilis, which was not included in the present investigation, seems to be somewhat different, however (18). The four chorismate synthases listed in Table V resemble each other very much in their cofactor specificities, kinetic properties, isoelectric points, and ph optima. Also, the structures of the enzymes must be similar as judged from the possession of common antigenic determinants as well as from sequence comparisons. In the prechorismate pathway, conservation ofthe amino acid sequences, although to a much lesser extent, has also been observed between the individual monofunctional E. coli enzymes and the corresponding domains of the fungal arom complex (9). These enzymes, therefore, appear to have originated from common ancestors, but presumably at some later stage during evolution the organizational forms of the pathway diverged. Enzymes were clustered to different degrees in bacteria, plants, fungi, and euglenoids. Enzyme organization also appears to be the major difference between the chorismate synthases of these organisms. Chorismate synthase can either be monofunctional, as in plants (26, 28) and in many bacteria (33), or it can be associated with a flavin reductase activity, as in N. crassa (13, 33) and B. subtilis (18). Chorismate synthase from E. gracilis was also found to be bifunctional. Whether or not the flavin reductase in E. gracilis is associated covalently with the chorismate synthase remains to be elucidated. As outlined in the introduction, a common feature of Euglena and higher fungi, in contrast with plants and bacteria, is the high degree to which different enzyme activities are aggregated. Chorismate synthase appears to fit into this scheme. ACKNOWLEDGMENT We wish to thank Professor J.R. Coggins (University of Glasgow, UK) for the kind gift of the strain E. coli AB2849/pGM62 and of a polyclonal rabbit antiserum directed against E. coli chorismate synthase. LITERATURE CITED 1. Ansorge W (1983) Fast visualization of protein bands by impregnation in potassium permanganate and silver nitrate. In D Stathakos, ed, Proceedings of Electrophoresis. W. de Gruyter, Berlin, pp Bagge P, Larsson C (1986) Biosynthesis of aromatic amino acids by highly purified spinach chloroplasts-compartmentation and regulation of the reactions. Physiol Plant 68: Bartlett PA, Maitra U, Chouinard PM (1986) Synthesis of "Iso- EPSP" and evaluation of its interaction with chorismate synthase. J Am Chem Soc 18: Bentley R (199) The shikimate pathway-a tree with many branches. Crit Rev Biochem Mol Biol 25: Biger P, San Pietro A (1967) Ferredoxin and cytochrome f in Euglena gracilis. Z Pflanzenphysiol 58: Byng GS, Whitaker RJ, Jensen RA (1985) Glyphosate as a probe of the pentafunctional arom protein of Euglena gracilis. Can J Bot 63:

9 CHARACTERIZATION OF CHORISMATE SYNTHASE Byng GS, Whitaker RJ, Shapiro CL, Jensen RA (1981) The aromatic amino acid pathway branches at L-arogenate in Euglena gracilis. Mol Cell Biol 1: Charles IG, Lamb HK, Pickard D, Dougan G, Hawkins AR (199) Isolation, characterization and nucleotide sequence of the aroc genes encoding chorismate synthase from Salmonella typhi and Escherichia coli. J Gen Microbiol 136: Coggins JR, Duncan K, Anton IA, Boocock MR, Chaudhuri S, Lambert JM, Lewendon A, Millar G, Mousdale DM, Smith DDS (1987) The anatomy of a multifunctional enzyme. Biochem Soc Trans 15: Dyer WE, Weaver LM, Zhao J, Kuhn DN, Weller SC, Herrmann KM (199) A cdna encoding 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase from Solanum tuberosum. J Biol Chem 265: Egan AF, Gibson F (197) Anthranilate synthase and anthranilate-5'-phosphoribosyl- I -pyrophosphate phosphoribosyl transferase (PR transferase) aggregate from Aerobacter aerogenes. Methods Enzymol 17: Fritz BJ, Ninnemann H (1985) Photoreactivation by triplet flavin and photoinactivation by singlet oxygen of Neurospora crassa nitrate reductase. Photochem Photobiol 41: Gaertner FH (1987) Chorismate synthase: a bifunctional enzyme in Neurospora crassa. Methods Enzymol 142: Gasser CS, Winter JA, Hironaka CM, Shah DM (1988) Structure, expression, and evolution of the Senolpyruvylshikimate- 3-phosphate synthase genes of petunia and tomato. J Biol Chem 263: Gibson F (197) Preparation of shikimic acid. Methods Enzymol 17: Hankins CN, Mills SE (1976) Anthranilate synthase-amidotransferase (combined). A novel form of anthranilate synthase from Euglena gracilis. J Biol Chem 251: Hankins CN, Mills SE (1977) A dimer of a single polypeptide chain catalyzes the terminal four reactions of the L-tryptophan pathway in Euglena gracilis. J Biol Chem 252: Hasan N, Nester EW (1978) Purification and properties of chorismate synthase from Bacillus subtilis. J Biol Chem 253: Hollinder H, Kiltz H-H, Amrhein N (1979) Interference of L-aaminooxy-,B-phenylpropionic acid with phenylalanine metabolism in buckwheat. Z Naturforsch Sect C Biosci 34: Homeyer U, Schultz G (1988) Activation by light of plastidic shikimate pathway in spinach. Plant Physiol Biochem 26: Knowles PF, Sprinson DB (197) Preparation of shikimate 5-phosphate. Methods Enzymol 17: Knowles PF, Levin JG, Sprinson DB (197) Preparation of 3- enolpyruvyl-shikimate 5-phosphate. Methods Enzymol 17: Laemmli UK (197) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: Lanzetta PA, Alvarez LJ, Reinach PS, Candia OA (1979) An improved assay for nanomole amounts ofinorganic phosphate. Anal Biochem 1: Morell H, Clark MJ, Knowles PF, Sprinson DB (1967) The enzymic synthesis of chorismic and prephenic acids from 3- enolpyruvylshikimic acid 5-phosphate. J Biol Chem 242: Mousdale DM, Coggins JR (1986) Detection and subcellular localization of a higher plant chorismate synthase. FEBS Lett 25: Patel VB, Giles NH (1979) Purification ofthe arom multienzyme aggregate from Euglena gracilis. Biochim Biophys Acta 567: Schaller A, Windhofer V, Amrhein N (199) Purification of chorismate synthase from a cell culture of the higher plant Corydalis sempervirens Pers. Arch Biochem Biophys 282: Schmidt CL, Danneel H-J, Schultz G, Buchanan BB (199) Shikimate kinase from spinach chloroplasts. Purification, characterization, and regulatory function in aromatic amino acid biosynthesis. Plant Physiol 93: Vogel HJ (1964) Distribution of lysine pathways among fungi: evolutionary implications. Am Nat 98: Weber HL, Buck A (197) Chorismate mutase from Euglena gracilis. Purification and regulatory properties. Eur J Biochem 16: Welch GR, Cole KW, Gaertner FH (1974) Chorismate synthase of Neurospora crassa: a flavoprotein. Arch Biochem Biophys 165: White PJ, Millar G, Coggins JR (1988) The overexpression, purification and complete amino acid sequence of chorismate synthase from Escherichia coli K12 and its comparison with the enzyme from Neurospora crassa. Biochem J 251: